One Year After AWS Savings Plans: What Have We Learned?

5 Min Read

One year ago, Amazon Web Services (AWS) rolled out a new discount program called Savings Plans, giving cloud managers a new option for saving money by making an advanced commitment. So, after one year of working with AWS Savings Plans, what have we learned? Here are three key takeaways from our experiences with thousands of customers who buy discounts from AWS.

Reserved Instances are not going away anytime soon

When AWS Savings Plans first hit the scene, many people compared Reserved Instances vs. Savings Plans and speculated that Reserved Instances would phase out, as Savings Plans offer the same discount rates on compute services, but with more flexibility. Let me explain.  

The discount rate for the two types of AWS Savings Plans maps to the discount rate for the two types of Reserved Instances.

  • Standard Reserved Instances and EC2 Savings Plans offer the highest discount (up to 72% off on-demand prices), but they're the most rigid—they can only be applied to EC2 instances in a specific family in a specific region.
  • Convertible Reserved Instances and Compute Savings Plans offer slightly fewer savings (up to 66% off on-demand prices), but they apply to any family or region. Reserved Instances must be manually converted, while Savings Plans will automatically be applied to any applicable compute service in any family or region.

aws savings plans vs reserved instances table

Because AWS Savings Plans offer the same discount rates as Reserved Instances with greater flexibility, some people speculated that this was the beginning of the end for Reserved Instances. And they may be right... eventually. But what we’ve learned over the past year is that Reserved Instances are not dead. Here’s why: 

  • AWS Savings Plans are still new. Many businesses continue to invest in Reserved Instances because they’re familiar with them.
  • AWS Savings Plans are sold differently than Reserved Instances. Instead of committing to a certain number of usage hours, you commit to spending a certain amount of money per hour on compute services. This shift will take a little time for businesses to adjust.
  • AWS Savings Plans only apply to compute services. At the moment, there are no Savings Plans that apply to database services.
  • There's no secondary market for Savings Plans. Businesses cannot sell unused Savings Plans back to the market like they can with Reserved Instances.
  • There are no capacity guarantees with Savings Plans. Unlike with Reserved Instances, you cannot guarantee capacity by assigning the Reserved Instance to a specific Availability Zone.

Read our in-depth article for a comparison of AWS Savings Plans vs. Reserved Instances.

Compute Savings Plans are more popular than EC2 Savings Plans

Even though EC2 Savings Plans offer the potential to save 6% more, Compute Savings Plans are winning the popularity contest. Why? Here’s the skinny: 

Seemingly, businesses are willing to forego a small amount of extra savings to have more flexibility. When Compute Savings Plans were first available, they only applied to EC2 and Fargate. Shortly thereafter, AWS added Lambda to the list. Cloud is a fast-moving space and AWS is constantly adding new services and expanding their offerings. Savings Plans are purchased on a one-year or three-year commitment timeline, and it’s reasonable to think that in that time, AWS will apply the Compute discount to more compute services.  

Compute Savings Plans are easier to manage too. While Convertible Reserved Instances can be exchanged for different types of instances, the process is manual and requires considerable oversight and usage monitoring. With Compute Savings Plans, discounts are applied automatically across all applicable components, significantly reducing management overhead.   

CloudHealth’s Compute Savings Plans Recommendations are simply the best out there. We break it down for you so you can quickly determine which Compute Savings Plan helps you meet your goal—whether you want to maximize savings, cover a certain amount of usage by discounts, or specify an hourly commitment amount in dollars.

aws compute savings plans cloudhealth dashboard recommendations

A best practice is to buy AWS Savings Plans at the master organization level

Like Reservations, AWS Savings Plans will “float” between linked accounts with an affinity for the purchasing account. Savings Plans are applied to the purchasing account's usage first, and then to other accounts' usage. This means that whenever possible, you should purchase Savings Plans from the top-level account—called the root account—in order to maximize the discount.

best practice for purchasing aws savings plans

Purchasing a Savings Plan from the root means you're more likely to utilize the entire discount. If the root organization doesn't spend its minimum monetary commitment, the discount will flow down to lower-level linked or consolidated accounts and be applied to compute resource usage there.

It's also important to note that that if you're purchasing both types of Savings Plans (and possibly Reserved Instances too), AWS will apply them in this order: 

  1. Standard Reserved Instance
  2. Convertible Reserved Instance
  3. EC2 Savings Plan
  4. Compute Savings Plan

As a whole, AWS Savings Plans have been a big hit in their first year and we expect adoption to continue to grow. CloudHealth has been providing tools to purchase, monitor, and adjust reservations since day one, and we will continue to support discount planning and management as these offers grow.

For more information on AWS Savings Plans and how to best utilize them, see our guide: The Ultimate Guide To AWS Savings Plans

Image
amber-gregorio-cloudhealth-bio
Amber Gregorio, Sr. Product Marketing Manager

Amber Gregorio is a Senior Product Marketing Manager at CloudHealth by VMware. She and her team own the development and implementation of the go-to-market strategy for the CloudHealth’s industry-leading multi- and hybrid cloud management software. Prior to joining CloudHealth, Amber worked in Product Marketing for an AI-powered customer engagement solution, and spent over 10 years in the financial services industry. Outside of work Amber enjoys weightlifting and competitive powerlifting.

We Think You Might Like These: